Two Channel AVHRR Discrimination
of Volcanic Clouds

Band subtraction technique

Thermal image data from two channels of the Advanced Very High Resolution Radiometer (AVHRR) were used in this study. Band 4 (10.3 to 11.3 um) minus band 5 (11.5 to 12.5 um) brightness temperature difference images are used to detect the volcanic cloud, and distinguish it from meteorological clouds. Volcanic clouds are known to have negative band 4 minus 5 brightness temperature differences (Prata, 1989a; Schneider et al., in press; Wen and Rose, 1994; Wen and Rose, online poster,), while meteorological clouds generally have positive brightness temperature differences (Yamanouchi et al, 1987). The images of the Mt. Spurr cloud (below) demonstrate the band subtraction technique. The image on the left is a band 4 image, and shows how difficult it is to distinguish the volcanic cloud. The image on the right is a band 4-5 image, and shows the volcanic cloud clearly.

For volcanic clouds, the magnitude of the negative brightness temperature difference is dependent upon the optical thickness of the cloud, the amounts of water, volcanic ash and sulfuric acid in the cloud, the size and size distribution of the particles in the cloud, and the temperature contrast between the cloud and the underlying surface (meteorological clouds, land or water) (Prata, 1989a; Wen and Rose, 1994).


Effect of cloud transparency on discrimination

Cloud opacity is an important factor in the discrimination of volcanic clouds using the band 4-5 technique. For opaque volcanic and meteorological clouds, all of the emitted thermal energy from the surface underlying the cloud is absorbed by the cloud. The radiance recorded by the satellite sensor is only that which is emitted by the cloud. At typical cloud temperaures, The band 4 radiance is greater than the band 5 radiance, which results in a positive band 4 minus 5 brightness temperature difference. Most of the earth's surface (water, land, snow, etc..) has a positive band 4 minus 5 temperature difference.

For transparent clouds, the radiance recorded by the satellite sensor is a combination of the radiance of the cloud, and the radiance of the underlying surface. For meteorological clouds, the absorption of the of energy from the underlying surface is greater in band 5 than in band 4, due to a larger refractive indices for water and ice particles in the band 5 region than in the band 4 region. This results in a positive band 4 minus 5 brightness temperature difference.

For transparent volcanic clouds, the absorption of energy from the underlying surface is greater in the band 4 region than the band 5 region, due to a larger refractive index for silicates in the band 4 region than in the band 5 region. This results in a negative band 4 minus 5 brightness temperature difference.


Evolution of the Volcanic Cloud Signal

The evolution of the volcanic cloud signal can be seen in a series of images of the August 19, 1992 volcanic cloud. Volcanic clouds imaged during eruption and shortly thereafter are optically thick and probably contain abundant water droplets and/or ice, which makes their spectral signal very much like a meteorological cloud. As the volcanic cloud disperses and becomes transparent, the spectral properties of the cloud change, first at the edge and then throughout. This produces a volcanic cloud signal that can be distinguished using a brightness temperature difference determined from thermal bands 4 and 5 of the AVHRR.

The progression from an opaque volcanic cloud to a transparent one has been observed in part in clouds from Galunggung (Prata, 1989b), Augustine (Holasek and Rose, 1991), Redoubt (Schneider and Rose, 1994), and Pinatubo, but this is the first instance where the full progression has been observed in one cloud.


Observations of the volcanic cloud from the August 19, 1992 eruption of Crater Peak Vent, Mt. Spurr.

The earliest image, collected about ninety minutes after the start of the eruption (upper left), shows a cold, circular volcanic cloud in band 4, and high thermal contrast between the volcanic cloud and the lower and warmer meteorological clouds. The band 4 minus 5 apparent brightness temperature difference algorithm does not work well on this image (upper right). The second image in this series was collected two hours later and shows a larger cloud in the band 4 image (lower left), and a fringing response from the band 4 minus 5 process (lower right).


The third image (upper left) was collected ninety minutes later. By this time the eruption had ended, and the fringe of negative band 4 minus 5 values had grown and encircled the entire volcanic cloud (upper right). The fourth image was collected eight hours later. The volcanic cloud is difficult to distinguish from the meteorological clouds in the band 4 image (lower left), but it is discriminated by the band 4 minus 5 operation (lower right), with the entire cloud showing negative brightness temperature difference values.


Long Distance Volcanic Cloud Tracking

The utility of the band 4 minus 5 operation to track clouds for days following an eruption was tested using two archived AVHRR data sets. The left figure is a composite of five images of the August volcanic cloud, and the right figure is a composite of nine images of the September volcanic cloud. Both clouds were tracked for more than eighty hours as the were transported thousands of kilometers. Because many factors affect the magnitude of the negative brightness temperature difference, the regions of the cloud with the greatest temperature difference do not necessarily have the highest concentration of particles.


References

Holasek, R.E. and W.I. Rose, Anatomy of 1986 Augustine Volcano eruptions as recorded by multispectral image processing of digital AVHRR weather satellite data, Bulletin of Volcanology, 53, 420-435, 1991.

Neal C.A., R.G. McGimsey, C.A. Gardner, M.L. Hardin, and C.J. Nye, Tephra-fall from the 1992 eruptions of Crater Peak, Mount Spurr, AK: A preliminary report on distribution, stratigraphy and composition, U.S. Geological Survey Bulletin, (ed.) T. Keith, in press.

Prata, A.J., Infrared radiative transfer calculations for volcanic ash clouds. Geophysical Research Letters, 16, 1293-1296, 1989a.

Prata, A.J., Observations of volcanic ash clouds in the 10-12 m window using AVHRR/2 data, International Journal of Remote Sensing, 10, 751-761, 1989b.

Schneider, D.J. and W.I. Rose, Observations of the 1989-90 Redoubt Volcano eruption clouds using AVHRR satellite imagery, Proceedings of the First International Symposium on Volcanic Ash and Aviation Safety, U.S. Geological Survey Bulletin 2047, (ed.) T. Casadevall, 405-418, 1994.

Schneider, D.J., W.I. Rose and L. Kelley, Tracking of 1992 Crater Peak/Spurr eruption clouds using AVHRR, U.S. Geological Survey Bulletin, (ed.) T. Keith, in press.

Wen S. and W.I. Rose, Retrieval of sizes and total masses of particles in volcanic clouds using AVHRR bands 4 and 5, Journal of Geophysical Research, 99, 5421-5431, 1994.

Wen S. and W.I. Rose, Retrieval of particle sizes and total masses in volcanic clouds using AVHRR bands 4 and 5, Poster presented at AGU 1993 Fall Meeting.

Yamanouchi, T., K. Suzuki and S. Kawaguchi, Detection of clouds in Antarctica from infrared multispectral data of AVHRR, Journal of the Meteorological Society of Japan, 65, 949-962, 1989.